Printer Friendly

Effects of different social and environmental conditions on established dominance relationships in crayfish.

Abstract. Like most social animals, crayfish readily form dominance relationships and linear social hierarchies when competing for limited resources. Competition often entails dyadic aggressive interactions, from which one animal emerges as the dominant and one as the subordinate. Once dominance relationships are formed, they typically remain stable for extended periods of time; thus, access to future resources is divided unequally among conspecifics. We previously showed that firmly established dominance relationships in juvenile crayfish can be disrupted by briefly adding a larger conspecific to the original pair. This finding suggested that the stability of social relationships in crayfish was highly context-dependent and more transient than previously assumed. We now report results that further identify the mechanisms underlying the destabilization of crayfish dominance relationships. We found that rank orders remained stable when conspecifics of smaller or equal size were added to the original pair, suggesting that both dominant and subordinate must be defeated by a larger crayfish in order to destabilize dominance relationships. We also found that dominance relationships remained stable when both members of the original pair were defeated by larger conspecifics in the absence of their original opponent. This showed that dominance relationships are not destabilized unless both animals experience defeat together. Lastly, we found that dominance relationships of pairs were successfully disrupted by larger intruders, although with reduced magnitude, after all chemical cues associated with earlier agonistic experiences were eliminated. These findings provide important new insights into the contextual features that regulate the stability of social dominance relationships in crayfish and probably in other species as well.

Introduction

Social animals aggressively compete for food, territory, shelter, and mates, because of the limited availability of resources (Wilson, 1975). In crustaceans, among other animals, competition leads to the formation of social ranks within social hierarchies, which help to determine the order of access to limited resources (Herberholz, 2014). After a social hierarchy (in groups) or dominance relationship (in pairs) is formed, aggressive interactions decrease, allowing individual animals to optimize their allocation of energy and time at a reduced risk of injury (Issa et al., 1999; Edwards and Herberholz, 2005; Herberholz et al., 2007).

Crayfish present an attractive model for the study of aggression and social hierarchies because of their ritualized, agonistic behavior (Huber et al., 2001; Kravitz and Huber, 2003). A pair of unacquainted crayfish will readily form a dominance relationship, even in the absence of resources other than space (Bovbjerg, 1953; Herberholz et al., 2007). In a pair of size-matched crayfish, formation of agonistic behaviors escalates to high-intensity levels (Schroeder and Huber, 2001), whereas fights between crayfish of markedly different sizes are often short and of low intensity (Pavey and Fielder, 1996; Figler et al., 1999). The early stages of an encounter between two crayfish include aggressive elements such as elevated body posture and raised claws. Fighting behavior includes interlocking claws, pushing the opponent, or striking the opponent with the claws, and offensive tail-flips (Bruski and Dunham, 1987; Herberholz et al., 2001). At some point, one animal breaks

off the fight by retreating or escaping with a rapid tail-flip (Issa et al., 1999; Herberholz et al., 2001). This sudden change in agonistic behavior signifies the formation of the dominance relationship, and identifies the newly emerged subordinate and dominant (Herberholz et al., 2003). The new dominant will continue to be aggressive and to initiate further interactions, whereas the new subordinate will retreat more readily to avoid its opponent (Issa et al., 1999; Goessmann et al., 2000).

Several factors determine whether a crayfish will emerge as the dominant or the subordinate. These include relative size (Pavey and Fielder, 1996), individual aggressiveness (Issa et al., 1999), previous residency (Peeke et al., 1995), and former winning or losing experiences (Goessmann et al., 2000). During fights, chemical cues play an important role in recognition of social status. Crayfish use unknown compounds in their urine to communicate their aggressive state, and blockage of urine release will change agonistic behavior (Zulandt Schneider and Moore, 2000; Zulandt Schneider et al., 2001). Moreover, exposure to chemical cues affects an opponent's behavior, causing it to act more dominantly when exposed to chemicals from subordinates, or act more submissively when exposed to chemicals from dominants (Martin et al., 2003; Bergman et al., 2005).

While the formation of social hierarchies has been studied extensively in various taxa, and the underlying behavioral and physiological factors thoroughly investigated, the interplay between contextual information and stability of social hierarchies is much less understood (Earley and Dugatkin, 2006). Chase et al. (2003) showed that stability and replication of dominance relationships in cichlid fish depend on the presence of other conspecifics. Social ranks are less stable and easier to reverse in pairs embedded in social contexts than in pairs tested in isolation. Similarly, Ewing (1972) showed that, in cockroaches, the organizational stability of social groups decreases as population density increases. More recently, Graham and Herberholz (2009) demonstrated how changes in context affect the stability of dominance relationships in crayfish. In a pair of juvenile crayfish with established social ranks, the temporary introduction of a larger conspecific disrupted the dominance relationship of the original pair. After the intruder was removed and the former dominants and former subordinates were allowed to reestablish their relationships, rank reversals were observed in half of the trials; that is, half of the original dominant animals became new subordinates and half of the original subordinates, new dominants. In a control group without intruders, no rank reversals were observed in any of the pairs. The Graham and Herberholz (2009) study reported that social ranks in crayfish can be transitive and reversible under certain social conditions; it showed that crayfish that won against an opponent and subsequently lost to another one can be defeated by conspecifics that lost their two previous fights.

However, that study also left several important questions unanswered. Since only larger intruders were used, it remained unclear whether the addition of a smaller or size-matched conspecific to a dominant-subordinate pair would be equally effective in destabilizing established social relationships. In addition, the rank reversals that were observed could have resulted from the single losing experience of the dominant (after the previous win), "resetting" the dominant to subordinate status, and thus having two new subordinates re-forming a dominance relationship with equal opportunity for each animal to obtain higher rank. Thus, whether rank reversals also occur after the dominant and subordinate are defeated in the absence of the original opponent remains untested. Moreover, these experiments were performed within an unchanged spatial and chemical environment. Therefore, it is unclear whether dominance relationships can be destabilized after sensory cues associated with the formation of the original relationship or with the defeat by the intruder are eliminated. In addition, the underlying behavioral mechanisms for the destabilization of established dominance relationships have not been identified. The current study addresses these open questions.

Materials and Methods

Basic research design

Juvenile red swamp crayfish Procambarus clarkii (Girard, 1852) were obtained from a commercial supplier (Atchafalaya Biological Supply, Raceland, LA), and kept in communal tanks. Animals in the study were sexually immature juveniles, and thus males and females were used randomly in the experiments. Prior to the experiment, the animals were individually isolated for one week in 10 X 15 X 8 cm (i.e., h X 1 X w) plastic tanks filled with gravel and water (to three quarters of the way to the top) to prevent any social experiences before testing. The animals were fed only one small-sized food pellet on the first day of isolation. Animals were kept at constant temperature (~21 [degrees]C) and under a 12:12 light-dark cycle.

The experimental design was adopted from a previous study (Graham and Herberholz, 2009). On the day of the experiment, animals were checked for missing appendages and molting. Only intact animals that had not molted within the last two days prior to the experiment were used. These animals were measured and marked with a nontoxic permanent marker for identification purposes. Two crayfish (focus animals) of the same length were placed in a 29 X 18 X 17 cm (h X 1 X w) plastic tank with 2 cm of gravel and deionized water, filled three quarters of the way to the top. The sides of the container were covered with white paper to minimize visual distractions. Each trial was done under normal fluorescent lighting. For the first 5 min, the animals were allowed to acclimate while an opaque plastic divider prevented visual or tactile contact between them.

Three experimental phases, each lasting 20 min, were recorded by a video camera positioned above the tank. At the beginning of the first phase (P1; 0-20 min), the divider was removed and the animals were allowed to interact and to establish a dominance relationship. At the beginning of the second phase (P2; 20-40 min), or the intruder phase--and depending on the experimental condition (see next section)--a third animal (intruder) was added to the original pair or to each member of the pair. The intruder(s) were then removed at the start of the third phase (P3; 40-60 min) so that the two focus animals had an opportunity to reestablish a dominance relationship. Transfers of crayfish were made using a fish net. Between each trial, the tanks were thoroughly rinsed with tap water. No animal was used in an experiment more than once, and no animal was injured during any of the trials.

Experimental design

In Experiment 1, intruders (length: 3.32 [+ or -] 0.14 cm, measured from rostrum to telson) that were smaller than the focus animals (4.[O.sub.3] [+ or -] 0.14 cm) were used. In each of the 12 trials, the two focus animals were size-matched; intruders were, on average, 17.7% [+ or -] 1.4% smaller than the focus animals. All three experimental phases (P1-P3) were conducted in the same tank (Fig. 1A). This experiment aimed to determine whether smaller intruders could elicit rank reversals in the original pair.

In Experiment 2, focus animals (3.68 [+ or -]0.18 cm) and intruders (3.66 [+ or -]0.16 cm) were size-matched in 12 trials. All three experimental phases (P1-P3) were conducted in the same tank (Fig. IB). The goal of this experiment was to determine whether same-sized intruders could elicit rank reversals in the original pair.

In Experiment 3, larger intruders were added to the tanks of each focus animal. In each of the 15 trials, the two focus animals were size-matched (3.63 [+ or -] 0.11 cm); intruders (4.22 [+ or -]0.14 cm) were, on average, 14.1% [+ or -] 1.6% larger than the focus animals. At the end of the first phase (P1), both members of the original pair were retrieved with a fish net and placed in two separate, new tanks of identical size, filled with gravel and fresh deionized water. Each tank was separated into two compartments by a divider, and each animal from the original pair was placed on one side in each tank. Both tanks were in view of the camera. A larger intruder was added to each tank on the other side of the divider. After 5 min, the divider was removed and agonistic interactions were recorded (P2). After 20 min (P3), both members of the original pair were placed in a single new tank, where they were allowed to acclimate for 5 min before their behavior was recorded for another 20 min (Fig. 1C).

The experiment was conducted to test whether separate defeat of both focus animals could lead to subsequent rank reversals. As a control, size-matched crayfish (3.70 [+ or -] 0.09 cm) were allowed to interact during P1; they were subsequently isolated in separate tanks for 20 min (P2) without adding a conspecific to each animal. They were reunited afterwards in a new tank and allowed to interact again for 20 min (P3). The control was conducted to see if transferring the animals twice, and temporary separation from the opponent, could alone cause distabilization of the dominance relationship.

In Experiment 4, intruders larger (4.26 [+ or -] 0.14 cm) than the focus animals (3.63 [+ or -] 0.10 cm) were used in 12 trials. The intruders were, on average, 14.8% [+ or -] 2.5% larger than the focus animals. The animals remained in the same tank throughout the first (P1) and second (P2) phases. Then, after removal of the intruder, at the beginning of the last phase (P3) the focus animals were moved into a new tank with gravel and fresh deionized water. The animals were allowed to acclimate for 5 min while separated by a barrier. After the acclimation period, the barrier was removed and the last phase (P3) began (Fig. ID). The goal of this experiment was to determine whether chemical cues released during the intruder period affected the stability of established dominant-subordinate relationships. As a control condition, size-matched animals (length: 3.61 [+ or -] 0.12 cm) were paired for P1 and P2 (20 min per phase) in the same tank without the addition of an intruder, and subsequently moved into a new tank for P3.

In Experiment 5, larger intruders (4.21 [+ or -] 0.15 cm) were added to the focus animals (3.65 [+ or -] 0.12 cm) in 14 trials. The intruders were, on average, 13.2% [+ or -] 1.7% larger than the focus animals. At the end of the first phase (P1), the animals were retrieved with a fish net and placed together in a new tank of identical size, filled with gravel and fresh deionized water. The animals were separated with an opaque plastic barrier and allowed to acclimate for 5 min before the barrier was removed; the intruder was then added, marking the beginning of the second phase (P2). At the beginning of the last phase (P3), the intruder was removed while the focus animals were left in the tank and allowed to interact for 20 min (Fig. IE). The experiment was performed to test whether chemical cues released during the initial formation of the dominance relationship affected its stability. We did not run another control for this experiment, because we had already tested the effects of transfer to a new tank after P1 as a control for Experiment 3 (see above).

Coding of behavior

After the trials were finished, videotaped recordings were analyzed, and agonistic behaviors of individual animals were measured separately for each trial phase (P1-P3). Trial phases were analyzed "blind," that is, in random order and by different experimenters to minimize any observer bias. The frequency of aggressive and submissive behavioral acts was counted for each animal. Aggressive acts were "approaches" (AP) and "attacks" (AT). An approach was defined as walking toward the other animal within one body length or turning around as if to face it (within one body length). An attack was moving toward the other animal with an elevated body posture and raised claw. Submissive acts were divided into "retreats" (RT) and "escapes" (ES). Retreat was walking away from an opponent, while escape was tail-flipping. A tail-flip is a rapid flexion of the abdomen that propels the crayfish away from its opponent. If an animal came close to another animal during the intruder phase while it was retreating or escaping from another opponent, the behavior was not counted as an approach. If an animal first walked backwards and then tail-flipped (as a response to an aggressive act by an opponent), the behavior was coded as an escape. If an animal tail-flipped multiple times in response to an attack or approach, each tail-flip was recorded as an escape.

The social relationship was assessed using a dominance index (DI), a measure of the relative levels of aggression, or the ratio of aggressive to total behaviors, ranging from 0 to 100%. The dominance index was calculated as follows: DI = (AP + 2X AT) / ((AP + 2XAT) + (RT + 2XES)) X 100. Because attacks are considered more aggressive acts than approaches, and escapes are considered more submissive acts than retreats, these behaviors were multiplied by two. The animal with the higher DI was assigned dominance status (DOM), and the animal with the lower DI was identified as the subordinate (SUB). Although there is ongoing debate in the literature about the relationship between aggression and dominance, and the definition of both (Drews, 1993), there are several reports across species showing a strong positive correlation between number of aggressive acts and dominance rank (e.g., Larson et al., 2006; Yurkovic et al., 2006; Surbeck et al., 2012). In addition, previous work in crayfish showed that our method of coding aggressive and submissive acts led to unambiguous identification of social rank, which was assessed by determining status-dependent priority of resource use (Herberholz et al., 2007) and by parallel expression of non-agonistic behaviors in dominants and subordinates (Herberholz et al., 2003).

Statistical analysis

All descriptive and analytical statistical analyses were performed using IBM SPSS Statistics, Version 23 (IBM, Armonk, NY). All values are presented as means [+ or -] standard deviation. Only non-parametric statistical tests (two-tailed) were used due to non-normality of most data, as established by one-sample Kolmogorov-Smirnov test. (Each statistical test is identified in the Results section, below.) We used Wilcoxon signed-rank tests for comparisons between animals that interacted with each other (related samples), and the Mann-Whitney U Test for comparisons between animals that had no interactions (unrelated samples). We used Fisher's exact test to compare frequency distributions.

Results

Experiments 1 and 2: smaller and same-sized intruders

Experiments 1 and 2 tested the possibility that an intruder crayfish, irrespective of its size, could disrupt established dominance relationships between a pair of crayfish. We found no reversal of dominance relationships in any of the trials in which a smaller intruder (Experiment 1 ) or same-sized intruder (Experiment 2) was used. Crayfish that became dominant in the first phase (P1) also dominated the original opponent in the third phase (P3), after both animals interacted with the intruder crayfish between the first and third phases for 20 min (P2).

Experiment 1 tested the effect of the addition of a smaller intruder. In the initial phase (P1), a dominance relationship was established, for which a significant difference (Wilcoxon signed-rank test: P [less than or equal to] 0.01) in dominance indices of dominants (97.9% [+ or -] 3.4%, n =12) and subordinates (20.5% [+ or -] 9.5%, n = 12) was found. The difference between the dominance indices of the original dominant and subordinate remained stable throughout the next two experimental periods (in P2, DOM = 97.9% [+ or -] 6.3%; SUB = 22.3% [+ or -] 14.1%; Wilcoxon signed-rank test: P [less than or equal to] 0.01; in P3, DOM = 100.0% [+ or -] 0.0%; SUB = 16.2% [+ or -] 13.9%; Wilcoxon signed-rank test: P [less than or equal to] 0.01; Fig. 2A). During the intruder phase (P2), the smaller intruders performed a similar number of aggressive acts (approaches and attacks) toward the dominants (7.5 [+ or -] 4.7) and the subordinates (7.0 [+ or -] 3.7, Mann-Whitney U test: P = 1.000). During this period, the dominants always defeated both their familiar size-matched opponents and the smaller intruders, while the subordinates always lost to the original opponents and defeated the smaller intruders in half of the trials.

This finding was reflected in the dominance indices of dominants, subordinates, and intruders during this period (Fig. 2B). In pairwise comparison, original dominants clearly dominated intruders (DOM = 98.7% [+ or -] 3.2%; Intruder (INT) = 26.3% [+ or -] 12.1%; Wilcoxon signed-rank test: P [less than or equal to] 0.01), whereas the dominance indices between original subordinates and intruders were similar and not significantly different (SUB = 54.8% [+ or -] 26.8%; INT = 50.5% [+ or -] 22.8%; Wilcoxon signed-rank test: P = 0.969). The subordinates defeated the smaller intruders in 6 trials (50%). In the last phase (P3), when subordinates interacted with initial dominants, their dominance indices (18.7% [+ or -] 15.0%, n = 6) were higher, but not significantly different from subordinates that had previously lost to the intruders (13.6% [+ or -] 13.5%, n = 6; Mann-Whitney U test: P = 0.518).

Experiment 2 tested the effects of a same-sized intruder on established social dominance relationships. In the first phase (P1), dominance, as determined by dominance indices, was firmly established between the members of the original pair (DOM = 95.4% [+ or -] 7.3%, n = 12; SUB = 23.8% [+ or -] 9.7%, n = 12; Wilcoxon signed-rank test: P [less than or equal to] 0.01, Fig. 3A). The dominance relationship remained stable throughout the next two experimental periods (in P2, DOM = 86.1% [+ or -] 28.1%; SUB = 30.2% [+ or -] 25.9%; Wilcoxon signed-rank test: P [less than or equal to] 0.01; in P3, DOM = 92.5% [+ or -]12.7%; SUB = 25.9% [+ or -] 17.9%; Wilcoxon signed-rank test: P [less than or equal to] 0.01; Fig. 3A). During the intruder phase (P2), intruders performed similar average numbers of aggressive acts toward the dominants (11.3 [+ or -] 7.9) and subordinates (12.0 [+ or -] 6.7; Mann-Whitney U test: P = 0.562). During this period, the same-sized intruder defeated the initial subordinate in 11 of 12 trials, and defeated the initial dominant in 5 of 12 trials. As a result, the dominance indices of original dominants and intruders were similar and not significantly different during this period (DOM = 63.4% [+ or -] 43.8%; INT = 55.8% [+ or -] 38.5%; Wilcoxon signed-rank test: P = 0.937), whereas the intruders clearly dominated the subordinates (SUB = 21.4% [+ or -] 16.1%; INT = 79.0% [+ or -] 24.1%; Wilcoxon signed-rank test: P [less than or equal to] 0.01; Fig. 3B). The difference in dominance indices between the two original crayfish was smaller during the last phase (P3) in trials when the dominant was defeated by the intruder (DOM = 83.9% [+ or -] 16.3%; SUB = 30.8% [+ or -] 18.4%, n = 5) compared to trials in which the intruder did not defeat the dominant animals (DOM = 98.6% [+ or -] 3.6%; SUB = 22.3% [+ or -] 18.0%, n = 1). Thus, the defeat of the dominant crayfish by the intruders in P2 decreased the dominance indices of the dominants, and caused an increase in the dominance indices of the subordinates when they reestablished their dominance relationship in P3. However, neither the change in the indices of the dominants (Mann-Whitney U test: P = 0.149) nor the change in the indices of the subordinates (Mann-Whitney U test: P = 0.432) was significant.

Experiment 3: separate defeat of dominants and subordinates

Experiment 3 tested the possibility that rank reversals were mediated by the defeat of the two original crayfish by the larger conspecifics, but did not require the presence of the original opponent. To test this theory, both members of the original pair (with established dominant and subordinate status) were defeated by larger conspecifics in separate tanks--and in the absence of the original opponent.

Interestingly, during this intruder period the larger intruders defeated all of the original subordinates, but only 80% of original dominants (12 of 15). Thus, a small number of dominants (3/15; 20%) were able to defeat the larger intruders during P2, which was never observed when the subordinates were present during P2. To be consistent with our methodology and to be able to compare across conditions, for our statistical analysis we used only the 12 pairs in which both animals were defeated by the larger intruder.

We found that dominance was firmly established during P1. Dominant animals had significantly higher dominance indices (97.2% [+ or -] 5.3%, n = 12) than subordinates (11.9% [+ or -] 12.6%, n = 12; Wilcoxon signed-rank test: P [less than or equal to] 0.01; Fig. 4A). Dominance indices remained significantly higher for original dominants than for original subordinates when the animals were reunited during P3 (DOM = 91.8% [+ or -] 15.2%; SUB = 24.9% [+ or -] 26.9%; Wilcoxon signed-rank test: P [less than or equal to] 0.01; Fig. 4A). However, a single rank reversal occurred (1/12; 8.3%) when one original subordinate and one dominant changed ranks during P3. During P2 (when each animal was now paired with a larger intruder in a separate tank), both the original dominants and the original subordinates experienced the same amount of aggression. Dominants experienced, on average, 17.5 [+ or -] 8.3 aggressive acts produced by the larger opponent, and subordinates experienced 16.2 [+ or -] 7.6 aggressive acts during this period, a non-significant difference (Mann-Whitney U test: P = 0.664). This finding was reflected in the overall dominance indices of intruders and original dominants (INT = 98.8% [+ or -] 3.5%; DOM = 9.9% [+ or -] 8.3%; Wilcoxon signed-rank test: P [less than or equal to] 0.01), as well as in the dominance indices of intruders and original subordinates (INT = 99.4% [+ or -] 1.5%; SUB = 8.6% [+ or -] 10.3%; Wilcoxon signed-rank test: P [less than or equal to] 0.01; Fig. 4B).

To test the effects of a 20-min separation period between P1 and P3, and to identify effects of multiple tank transfers, we performed the following control experiment: after P1, dominants and subordinates of equal sizes were isolated for 20 min in separate tanks, with no other animal present, and then reunited in a new tank for P3. We found that this condition had no effect on the stability of previously established rank orders. Animals that obtained dominance or subordinate status in P1 (DOM = 97.1% [+ or -] 7.0%; SUB = 11.6% [+ or -] 10.2%; Wilcoxon signed-rank test: P [less than or equal to] 0.01) remained at their respective ranks in P3 (DOM = 98.0% [+ or -] 4.8%; SUB = 15.7% [+ or -] 9.9%, n = 12; Wilcoxon signed-rank test: P [less than or equal to] 0.01; Fig. 4C). Moreover, the single rank reversal observed in the experimental group was not significantly different from the control (Fisher's exact test: P = 1.000).

Experiments 4 and 5: spatial and chemical cues

Experiment 4 tested the effects of spatial and chemical cues associated with winning and losing experiences during the intruder phase. As before, the two animals readily formed a firm dominance relationship during P1, as documented by the significant difference in their dominance indices (DOM = 99.7% [+ or -] 1.1%, n = 12; SUB = 15.2% [+ or -] 8.9%, n = 12; Wilcoxon signed-rank test: P [less than or equal to] 0.01). When the original animals were reunited in a new tank after interacting with the intruder, rank reversals occurred in 42% of all pairings (5/12). In addition, the dominance indices of the original dominants (65.8% [+ or -] 42.4%) and original subordinates (45.8% [+ or -] 41.3%) were no longer significantly different (Wilcoxon signed-rank test: P = 0.209; Fig. 5A). During P2, the dominance indices of both members of the original pair remained significantly different (DOM = 52.6% [+ or -] 32.5%; SUB = 14.0% [+ or -] 11.9%; Wilcoxon signed-rank test: P < 0.05), and the intruders performed a similar average number of aggressive acts toward the dominants (26.3 [+ or -] 15.3) and subordinates (21.7 [+ or -] 18.7; Mann-Whitney U test: P = 0.093). The intruders clearly dominated both the original dominants (INT = 95.5% [+ or -] 8.5%; DOM = 14.9% [+ or -] 12.9%; Wilcoxon signed-rank test: P [less than or equal to] 0.01) and the original subordinates (INT = 96.6% [+ or -] 10.1%; SUB = 3.9% [+ or -] 3.3%; Wilcoxon signed-rank test: P [less than or equal to] 0.01; Fig. 5B).

In the control condition, 12 pairs of size-matched crayfish were allowed to interact with each other for 40 min (P1+P2) before being moved to a new tank. In these pairs, in which no intruder was added during P2, no rank reversals occurred, and the dominance indices of dominants remained significantly higher than in subordinates over all three periods (in P1, DOM = 97.1% [+ or -] 5.6%; SUB = 28.5% [+ or -] 14.5%; Wilcoxon signed-rank test: P [less than or equal to] 0.01; in P2, DOM = 100% [+ or -] 0%; SUB = 26.3% [+ or -] 19.6%; Wilcoxon signed-rank test: P [less than or equal to] 0.01; in P3, DOM = 95.3% [+ or -] 9.8%; SUB = 33.0% [+ or -] 18.5%; Wilcoxon signed-rank test: P [less than or equal to] 0.01; n = 14; Fig. 5C). The number of rank reversals observed in the experimental group was significantly higher than in the control (Fisher's exact test: P [less than or equal to] 0.05).

Experiment 5 tested the effects of spatial and chemical cues associated with winning and losing experiences during the initial dominance formation. In P1, a dominance relationship was established; the dominance indices of the dominants (94.6% [+ or -] 6.5%, n = 14) were significantly higher than those of the subordinates (15.9% [+ or -] 13.6%, n = 14; Wilcoxon signed-rank test: P [less than or equal to] 0.01). The animals were then transferred into a new tank to interact with the intruder. After the intruder was removed (P3), the original pair remained in the tank and reestablished their dominance relationships. Under these conditions, rank reversals occurred in 36% of all dominant-subordinate pairs (5/14), and the dominance indices of the initial dominants (60.5% [+ or -] 38.1%) and initial subordinates (45.1% [+ or -] 37.3%) were no longer significantly different (Wilcoxon signed-rank test: P = 0.510; Fig. 6A). Interestingly, during P2, after animals were transferred into a new tank, the dominance indices of both members of the original pair were no longer significantly different (DOM = 48.2% [+ or -] 31.6%; SUB = 32.3% [+ or -] 24.9%; Wilcoxon signed-rank test: P = 0.209). However, the intruders performed a similar average number of aggressive acts toward the dominants (28.2 [+ or -] 20.5) and subordinates (21.8 [+ or -] 16.9; Mann-Whitney U test: P = 0.269), and they clearly dominated both the original dominants (INT = 96.1% [+ or -] 4.9%; DOM = 19.9% [+ or -] 14.3%; Wilcoxon signed-rank test: P [less than or equal to] 0.01) and the original subordinates (INT = 97.8% [+ or -] 3.5%; SUB = 13.1% [+ or -] 16%; Wilcoxon signed-rank test: P [less than or equal to] 0.01; Fig. 6B). The number of rank reversals in this group was significantly higher than the number of rank reversals observed in any of the control pairs (Fisher's exact test: P [less than or equal to] 0.05).

The underlying mechanisms of the observed destabilization of dominance relationships are still elusive. To better understand the behavioral patterns that promote rank reversals, we performed preliminary data analysis, quantifying aggressive and submissive acts displayed by dominants and subordinates in trials that did or did not lead to rank reversals (Table 1). We pooled the data from Experiments 4 and 5, since the number of reversals in both conditions differed significantly from our controls (Fisher's exact test: P [less than or equal to] 0.05), but not from each other (Fisher's exact test: P = 0.698). We found no differences in P1 (the time when the original pairs formed their dominance relationships) for dominance indices (DI) of dominants and subordinates between the two later outcomes. In trials that did not feature reversals of dominance relationships in P3 (non-reversals; n = 18), the DIs of dominants (97.7% [+ or -] 3.5%) were not significantly different from the DIs of dominants (95.5% [+ or -] 7.1%) of trials with later reversals (reversal; n = 10; MannWhitney U test: P = 0.655). Likewise, the DIs of subordinates in P1 in non-reversal trials (15.2% [+ or -] 11.8%) were not different from DIs of subordinates in reversal trials (16.9% [+ or -] 10.8%; Mann-Whitney U test: P = 0.655). This finding showed that both members of the pair went into the subsequent stages (P2, P3) without any obvious predisposition to status reversals.

However, in P2, the intruder period, some interesting differences started to emerge. While dominants (10.8 [+ or -] 6.8) performed a significantly higher number of aggressive acts (approaches and attacks) than subordinates (4.4 [+ or -] 4.7) in non-reversal trials (Wilcoxon signed-rank test: P [less than or equal to] 0.01), there was no significant difference (Wilcoxon signed-rank test: P = 0.173) in the number of aggressive acts performed by the dominants (9.3 [+ or -] 6.2) and subordinates (5.7 [+ or -] 5.8) in reversal trials (Table 1). In addition, in non-reversal trials dominants produced, on average, twice as many, and significantly more, aggressive acts toward the subordinates than the subordinates toward the dominants (DOM vs. SUB = 6.0 [+ or -] 3.9; SUB vs. DOM = 3.0 [+ or -] 3.2; Wilcoxon signed-rank test: P [less than or equal to] 0.05). In reversal trials, however, the difference in aggression displayed by dominants and subordinates was no longer significant (DOM vs. SUB = 5.1 [+ or -] 3.0; SUB vs. DOM = 3.5 [+ or -] 4.1; Wilcoxon signed-rank test: P = 0.398). Moreover, subordinates produced a significantly higher number of submissive acts (retreats and escapes) in response to dominants' attacks and approaches in non-reversal trials (10.0 [+ or -] 5.3) than did dominants in response to subordinates' aggressive behaviors (5.1 [+ or -] 4.7; Wilcoxon signed-rank test: P [less than or equal to] 0.01). This pattern changed in reversal trials when dominants and subordinates produced a very similar and not statistically different number of submissive acts in response to the opponents' aggression (DOM vs. SUB = 7.2 [+ or -] 5.7; SUB vs. DOM = 7.7 [+ or -] 3.8; Wilcoxon signed-rank test: P = 0.722).

Discussion

We previously showed that changes in social context can disrupt social hierarchies in crayfish. The temporary addition of a larger intruder to a pair of crayfish with established dominance relationships weakened and reversed these relationships (Graham and Herberholz, 2009). While the presence of the larger intruder caused both animals of the initial pair to perform a higher number of aggressive acts, both animals were always defeated by the larger intruder (Graham and Herberholz, 2009). Based on these results, in the current study we explored other contextual features that could affect the stability of dominance relationships.

We found that intruders of smaller or equal size were ineffective in causing any status reversals in pairs with established ranks. This showed that social relationships were not destabilized, and rank reversals were not mediated simply by agonistic interactions with an additional conspecific. Instead, for rank reversals to occur, the intruder has to be larger than both members of the original pair, which gives the intruder an agonistic advantage leading to defeats of both the dominants and subordinates. In experiments with smaller intruders, all dominants defeated the intruder. However, only some subordinates defeated the intruder while others lost despite their size advantage. Losing to a smaller conspecific is rarely observed in socially naive crayfish; however, losing (or winning) is known to decrease (or increase) agonistic success in future interactions, a phenomenon known as "winner-loser effect" across many taxa, including crayfish (Bergman et al., 2005; Hsu et al., 2006).

Thus, the observed losses of some subordinates to smaller intruders could be a consequence of the previous losing experience (to the dominant); since not all subordinates lost to the smaller intruder, however, differences in individual fighting success were likely influenced by the behavior of the dominants toward the subordinates during the intruder period. Interestingly, subordinates that won against the intruder were more successful (i.e., had higher dominance indices) in subsequent interactions with their former dominant opponent. Thus, the winning experience against the intruder may have empowered the original subordinates, which led to more agonistic success ("winner effect") when both animals reestablished their dominance relationship. In experiments with same-sized intruders, almost half of the former dominants were defeated by the intruder, which was surprising, given that dominants had a recent winning experience and should, therefore, be successful in defeating opponents of equal size. This finding strongly suggested that the presence of the subordinates during interactions with the intruder impaired the agonistic success of the dominants in these trials. It seems possible that the dominants struggled because they had to divide their aggression between the two opponents; in other words, they had to try both to defeat the intruder and contain the subordinate. Consequently, dominants that lost to the intruder also showed a decrease in dominance indices when they were facing the initial subordinates after the intruder period ("loser effect"). Conversely, subordinates had higher overall dominance indices after the intruder period than in their first fights, despite the fact that all but one subordinate was defeated by the same-sized intruder. Although no status reversals were found in these trials, our results imply that the defeat of the dominant animal by the same-sized intruder left traces in both members of the original pair. That is, dominants became less aggressive afterwards, and subordinates became more aggressive after they had witnessed the defeat of the dominants (see below).

When the initial dominant and subordinate were separated--after they first established a dominance relationship and each animal was paired individually with a larger conspecific--we observed only one rank reversal; the differences in dominance indices between the original opponents remained significant when they reestablished their relationship. Thus, the original dominance relationships remained stable. This was a surprising and important result, because it suggested that the loss experienced by the dominant against the intruder was alone insufficient to robustly produce rank reversals. It further supports the notion that dominance relationships are destabilized only when the dominant is defeated in the presence of the subordinate.

Although "bystander" effects have been described at both the empirical and theoretical levels in a number of species, including crayfish, the results are somewhat inconclusive. These prior studies primarily tested the effects of eavesdropping on fights rather than the effect of physical presence during the fights (Johnstone, 2001; Earley and Dugatkin, 2002; Peake et al., 2006; Aquiloni et al., 2008; Earley, 2010). Nonetheless, these experiments showed that animals witnessing other members of their species fight used this information to adjust their own behavior toward these conspecifics, sometimes by increasing aggression (Clotfelter and Paolino, 2003). Thus, subordinates that are present during the defeat of the dominants have ample opportunity to collect information related to the dominants' degradation, which may empower them to challenge the dominants in subsequent fights. How the subordinates evaluate the dominants' weakness is currently unknown, but will be a focus of our future research.

Importantly, in these trials with separate intruders a few dominant animals were able to defeat the larger intruder when facing them alone, which was probably mediated by the recent win they had experienced over the subordinates. We never observed dominants defeating larger intruders in the presence of subordinates; and some dominants lost fights against same-sized intruders under these conditions, which must have eliminated any advantage from the previous win. This provides further support for the idea that immediate agonistic success of the dominant is inhibited by the physical presence of the subordinate when facing another opponent (intruder), and that future agonistic success is also lessened when the dominant defends its status against the original opponent following the intruder period. Taken together, these results show that both dominants and subordinates are affected by the shared experience of being defeated by the intruders.

Chemical cues released during fights play an important role in the formation and maintenance of social hierarchies in crayfish. Crayfish release urine into the water to communicate their status; such signals strongly affect the behavior of the opponent (Bergman and Moore, 2001, 2005; Horner et al., 2008). We found that chemical cues in the water were not essential to promote status reversals. We changed the chemical environment before and after a larger intruder was added to the original pair, and observed multiple rank reversals in both cases. Thus, chemical cues left in the tank from the original fights, or from agonistic interactions with the intruder, are not necessary for the destabilization of dominance relationships. However, the observed 36% of reversals when the animals were transferred into a new tank before they encountered the intruder, and 42% of reversals when they were placed in new tanks after their interactions with the intruder, are lower than the percentage of rank reversals (50%) we observed in an earlier study, where animals were left in the same tank throughout the entire experimental period (Graham and Herberholz, 2009). Hence, the lowest number of rank reversals occurred when chemical cues were eliminated immediately after the first fights, and the highest percentage of rank reversals were observed when the chemical environment was unchanged throughout the experiment. However, in our last experiment, when we transferred the original pairs into new tanks for the intruder period, dominance indices of dominants and subordinates (measured against each other) were no longer significantly different. This finding implied that loss of chemical (and possibly spatial) information from the initial formation of the dominance relationship started to affect rank stability. Thus, we must consider the possibility that chemical cues released into the water play a role in regulating the stability of dominance relationships. To clarify the interplay between chemical cues and dominance stability, future experiments could be designed in which chemosensors are ablated and/or urine release is blocked during fights with the intruder.

Our analysis of reversal versus non-reversal fights suggests that important behavioral changes take place during the intruder period, changes that are predictive of the future stability of the dominance relationship. Rank reversals are signified by a reduction in aggression and an increase in submission by the original dominants, and by the opposite behavioral pattern in the original subordinates during P2. However, since our analysis was performed with a relatively small sample size, and it combined data of slightly different conditions (i.e., timing of transfers into new tanks), future work should focus on a large group that is tested under equal conditions to further elucidate the mechanisms that destabilize social relationships in crayfish.

To the best of our knowledge, the experimental design applied to our study has not been tested in other species, a situation that limits comparison across taxa. However, it seems likely that similar effects are taking place in other social species, including humans, where degradation of a superior in the presence of a subordinate may leave traces that destabilize hierarchical order and social cohesion. Thus, our study may inform work in other social, non-human animals as well as in humans, leading to a better understanding of the factors that determine social stability.

The added value of crayfish as a model system for investigating these factors stems from their suitability for uncovering the sensory and neurochemical underpinnings. It seems feasible to unscramble the individual roles of different sensory cues and their importance for destabilizing dominant-subordinate relationships (which apparently takes its course during the intruder period). For example, it has been shown that an aggressive posture displayed by a future opponent can quickly increase heart rate in crayfish (Listerman et al., 2000). Although likely to be used in conjunction with other sensory cues (e.g., chemical) during aggressive encounters, this result indicates that visual cues can alter physiological state, which, in turn, could mediate changes in fighting behavior.

In addition, new, exciting avenues present themselves to measure the role of monoaminergic neuromodulators in regulating the stability of dominance relationships (Herberholz, 2013).

Crayfish have long been a highly productive model with which to investigate the neural and neurochemical basis of aggression and dominance (Yeh et al., 1996; Edwards and Kravitz, 1997; Huber et al., 2001); the rapid reversal of dominance status observed in our study was previously described after claw loss in lobsters (Lang et al., 1977). However, whether claw loss and the corresponding dominance reversal are paralleled by monoaminergic changes remains to be determined. Both serotonin and octopamine affect agonistic behaviors and fighting success in lobsters and crayfish, but the effects of these monoamines vary widely depending on social context, investigated species, and procedure of drug application (Livingstone et al., 1980; McRae, 1997; Tiemey and Mangiamele, 2001; Momohara et al., 2013). Thus, future work can focus on identifying the neurochemistry underlying status reversals. Together, these results will greatly better our understanding of dominance hierarchies and social behaviors generally.

Acknowledgments

We thank Ms. Kelly Rollman for help with conducting some of the experiments and data analysis.

Literature Cited

Aquiloni, L., M. Bufic, and F. Gherardi. 2008. Crayfish females eavesdrop on fighting males before choosing the dominant mate. Curr. Biol, 18: R462-R463.

Bergman, D. A., and P. A. Moore. 2001. The olfactory role in the fight dynamics of the crayfish. Orconectes rusticus. Am. Zool. 41: 1390.

Bergman, D. A., and P. A. Moore. 2005. Prolonged exposure to social odours alters subsequent social interactions in crayfish (Orconectes rusticus). Anim. Behav. 70: 311-318.

Bergman, D. A., C. P. Kozlowski, J. C. Mclntyre, R. Huber, A. G. Daws, and P. A. Moore. 2003. Temporal dynamics and communication of winner-effects in the crayfish, Orconectes rusticus. Behaviour 140: 805-825.

Bergman, D. A., A. L. Martin, and P. A. Moore. 2005. Control of information flow through the influence of mechanical and chemical signals during agonistic encounters by the crayfish, Orconectes rusticus. Anim. Behav. 70: 485-496.

Bovbjerg, R. V. 1953. Dominance order in the crayfish Orconectes virilis (Hagen). Physiol. Zool. 26: 173-178. Bruski, C. A., and D. W. Dunham. 1987. The importance of vision in agonistic communication of the crayfish Orconectes rusticus. I: An analysis of bout dynamics. Behaviour 1[O.sub.3]: 83-107.

Chase, I. D., C. Tovey, and P. Murch. 2003. Two's company, three's a crowd: differences in dominance relationships in isolated versus socially embedded pairs offish. Behaviour 140: 1193-1217.

Clotfelter, E. D., and A. D. Paolino. 2003. Bystanders to contests between conspecifics are primed for increased aggression in male fighting fish. Anim. Behav. 66: 343-347.

Drews, C. 1993. The concept and definition of dominance in animal behaviour. Behaviour 125: 283-313.

Earley, R. L. 2010. Social eavesdropping and the evolution of conditional cooperation and cheating strategies. Philos. Trans. R. Soc. Lond. B Biol. Sci. 365: 2675-2686.

Earley, R. L., and L. A. Dugatkin. 2002. Eavesdropping on visual cues in green swordtail (Xiphophorus helleri) fights: a case for networking. Proc. R. Soc. Lond B Biol. Sci. 269: 943-952.

Earley, R. L., and L. A. Dugatkin. 2006. Merging social hierarchies: effects on dominance rank in male green swordtail fish (Xiphophorus helleri). Behav. Process. 73: 290-298.

Edwards, D. H., and J. Herberholz. 2005. Crustacean models of aggression. Pp. 38-61 in The Biology of Aggression, R. J. Nelson, ed. Oxford University Press, Oxford.

Edwards, D. H., and E. A. Kravitz. 1997. Serotonin, social status and aggression. Curr. Opin. Neurobiol. 7: 812-819.

Ewing, L. S. 1972. Hierarchy and its relation to territory in the cockroach [N.sub.a]uphoeta cinerea. Behaviour 42: 152-174.

Figler, M. H., H. M. Cheverton, and G. S. Blank. 1999. Shelter competition in juvenile red swamp crayfish (Procamharus clarkii): the influences of sex differences, relative size, and prior residence. Aquaculture 178: 63-75.

Goessmann, C, C. Hemelrijk, and R. Huber. 2000. The formation and maintenance of crayfish hierarchies: behavioral and self-structuring properties. Behav. Ecol. Sociobiol. 48: 418-428.

Graham, M. E., and J. Herberholz. 2009. Stability of dominance relationships in crayfish depends on social context. Anim. Behav. 77: 195-199.

Herberholz, J. 2013. Serotonergic modulation of aggression. Pp. 27-51 in Serotonin: Biosynthesis, Regulation and Health Implications, F. S. Hall, ed. NOVA Science Publishers, Hauppauge, NY.

Herberholz, J. 2014. Neurobiology of social status in crustaceans. Pp. 457-483 in The [N.sub.a]tural History of the Crustacea, Vol. 3: Crustacean Nervous Systems and Their Control of Behavior, C. Derby and M. Thiel, eds. Oxford University Press, New York.

Herberholz, J., F. A. Issa, and D. H. Edwards. 2001. Patterns of neural circuit activation and behavior during dominance hierarchy formation in freely behaving crayfish. J. Neurosci. 21: 2759-2767.

Herberholz, J., M. M. Sen, and D. H. Edwards. 2003. Parallel changes in agonistic and non-agonistic behaviors during dominance hierarchy formation in crayfish. J. Comp. Physiol. A Sens. Neural Behav. Physiol. 189: 321-325.

Herberholz, J., C. McCurdy, and D. H. Edwards. 2007. Direct benefits of social dominance in juvenile crayfish. Biol. Bull. 213: 21-27.

Horner, A. J., M. Schmidt, D. H. Edwards, and C. D. Derby. 2008. Role of the olfactory pathway in agonistic behavior of crayfish, Procamharus clarkii. Invertebr. Neurosci. 8: 11-18.

Hsu, Y., R. L. Earley, and L. L. Wolf. 2006. Modulation of aggressive behaviour by fighting experience: mechanisms and contest outcomes. Biol. Rev. Camb. Philos. Soc. 81: 33-74.

Huber, R., J. B. Panksepp, Z. Yue, A. Delago, and P. Moore. 2001. Dynamic interactions of behavior and amine neurochemistry in acquisition and maintenance of social rank in crayfish. Brain Behav. Evol. 57: 271-282.

Issa, F. A., D. J. Adamson, and D. H. Edwards. 1999. Dominance hierarchy formation in juvenile crayfish Procamharus clarkii. J. Exp. Biol. 202: 3497-3506.

Johnstone, R. A. 2001. Eavesdropping and animal conflict. Proc. Natl. Acad. Sci. USA 98: 9177-9180.

Kravitz, E. A., and R. Huber. 2003. Aggression in invertebrates. Curr. Opin. Neurobiol. 13: 736-743.

Lang, F., C. K. Govind, W. J. Costello, and S. I. Greene. 1977. Developmental neuroethology: changes in escape and defensive behavior during growth of the lobster. Science 197: 682-685.

Larson, E. T., D. M. O'Malley, and R. H. Melloni, Jr. 2006. Aggression and vasotocin are associated with dominant-subordinate relationships in zebrafish. Behav. Brain Res. 167: 94-102.

Listerman, L. R., J. Deskins, H. Bradacs, and R. L. Cooper. 2000. Heart rate within male crayfish: social interactions and effects of 5-HT. Comp. Biochem. Physiol. A 125: 251-263.

Livingstone, M. S., R. M. Harris-Warrick, and E. A. Kravitz. 1980. Serotonin and octopamine produce opposite postures in lobsters. Science 208: 76-79.

Martin, A. L., D. A. Bergman, and P. A. Moore. 2003. Social communication in crayfish: a simultaneous recording of urine signals and flow fields during agonistic encounters, lntegr. Comp. Biol. 43: 854. [Abstract].

McRae, T. 1997. On the postural effects induced in female Cherax destructor (Clark) by serotonin and octopamine. Freshw. Crayfish 11: 293-298.

Momohara, Y., A. Kanai, and T. [N.sub.a]gayama. 2013. Aminergic control of social status in crayfish agonistic encounters. PLoS One 8: e74489.

Pavey, C. R., and D. R. Fielder. 1996. The influence of size differential on agonistic behavior in the freshwater crayfish, Cherax cuspidatas (Decapoda: Parastacidae). J. Zool. 238: 445-457.

Peake, T. M., R. J. Matos, and P. K. McGregor. 2006. Effects of manipulated aggressive 'interactions' on bystanding male fighting fish, Betta splendens. Anim. Behav. 72: 1013-1020.

Peeke, H. V. S., J. Sippel, and M. H. Figler. 1995. Prior residence effects in shelter defense in adult signal crayfish (Pacifastacus leniusculus (Dana)): results in same-and mixed-sex dyads. Crustaceana 68: 873-881.

Schroeder, L., and R. Huber. 2001. Fight strategies differ with size and allometric growth of claws in crayfish, Orconectes rusticus. Behaviour 138: 1437-1449.

Surbeck, M., T. Deschner, G. Schubert, A. Weltring, and G. Hohmann. 2012. Mate competition, testosterone and intersexual relationships in bonobos. Pan paniscus. Anim. Behav. 83: 659-669.

Tierney, A. J., and L, A. Mangiamele. 2001. Effects of serotonin and serotonin analogs on posture and agonistic behavior in crayfish. J. Comp. Physiol. A 187: 757-767.

Wilson, E. O. 1975. Sociobiology: The New Synthesis. Harvard University Press, Cambridge, MA.

Yeh, S.-R., R. A. Fricke, and D. H. Edwards. 1996. The effect of social experience on serotonergic modulation of the escape circuit of crayfish. Science 271: 366-369.

Yurkovic, A., O. Wang, A. C. Basu, and E. A. Kravitz. 2006. Learning and memory associated with aggression in Drosophila melanogaster. Proc. Natl. Acad. Sci. USA 1[O.sub.3]: 17519-17524.

Zulandt Schneider, R. A., and P. A. Moore. 2000. Urine as a source of conspecific disturbance signals in the crayfish Procambarus clarkii. J. Exp. Biol. 203: 765-771.

Zulandt Schneider, R. A., R. Huber, and P. A. Moore. 2001. Individual and status recognition in the crayfish, Orconectes rusticus: the effects of urine release on fight dynamics. Behaviour 138: 137-153.

JENS HERBERHOLZ (1,2*), MATTHEW E. SWIERZBINSKI (1,2), AND JULIANE M. BIRKE (1)

(1) Department of Psychology and (2) Neuroscience and Cognitive Science Program, University of Maryland, College Park, Maryland 20742

Received 31 December 2015; accepted 15 March 2016.

(*) To whom correspondence should he addressed. E-mail: jherberh@umd.edu

Table 1
Aggressive and submissive behaviors displayed by crayfish in reversal
and non-reversal trials

                       P2 (Intruder Period)
                       Aggression (total)
                DOM                SUB         P

Non-REV  10.8 [+ or -] 6.8  4.4 [+ or -] 4.7  <0.01
REV       9.3 [+ or -] 6.2  5.7 [+ or -] 5.8   0.173

                      P2 (Intruder Period)
                        Aggression (OPP)
         DOM:SUB               SUB:DOM           P

Non-REV  6.0 [+ or -] 3.9  3.0 [+ or -] 3.2  <0.05
REV      5.1 [+ or -] 3.0  3.5 [+ or -]4.1    0.398

                        P2 (Intruder Period)
                          Submission (OPP)
              DOM:SUB           SUB:DOM            P

Non-REV  5.1 [+ or -]4.7   10.0 [+ or -] 5.3  [less than or equal to]
                                               0.01
REV      7.2 [+ or -] 5.7   7.7 [+ or -] 3.8   0.722

Average number [+ or -] standard deviation of aggressive (attack,
approach) and submissive (escape, retreat) behavioral acts produced by
dominants (DOM) and subordinates (SUB) during the intruder period in
trials that remained stable (Non-REV) and in trials that later (in
Phase 3 (P3)) reversed (REV). The Aggression (total) column shows the
average aggressive acts against all opponents (e.g., dominants against
subordinates and intruders); Aggression (OPP) and Submission (OPP) show
the average number of behaviors against the original opponent only
(e.g., dominants against subordinates). Values are means [+ or -] SD.
P-values are based on Wilcoxon signed-rank tests.
COPYRIGHT 2016 University of Chicago Press
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2016 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Herberholz, Jens; Swierzbinski, Matthew E.; Birke, Juliane M.
Publication:The Biological Bulletin
Article Type:Report
Date:Apr 1, 2016
Words:8868
Previous Article:The molting biomarker metabolite N-acetylglucosamino-l,5-lactone in female urine of the helmet crab Telmessus cheiragonus.
Next Article:No effects and no control of epibionts in two species of temperate pycnogonids.
Topics:

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters